A novel approach for APT attack detection based on an advanced computing.

Area of Science:

• Cybersecurity
• Network Security
• Machine Learning

Background:

• Advanced Persistent Threats (APTs) pose a significant risk to network security.
• Existing APT detection methods often struggle with accuracy and false predictions.
• Effective behavior profiling is crucial for identifying sophisticated cyber-attacks.

Purpose of the Study:

• To propose a novel approach for enhancing Advanced Persistent Threat (APT) detection.
• To build and analyze behavior profiles of APT attacks in network traffic.
• To improve the accuracy and reduce false predictions in APT detection.

Main Methods:

• Utilized a combination of Bidirectional Long Short-Term Memory (Bi) and Attention (A) for behavior profile construction.
• Employed a Dynamic Graph Convolutional Neural Network (DGCNN) for feature extraction and classification of APTs.
• Developed the BiADG model integrating these components for comprehensive APT analysis.

Main Results:

• The proposed BiADG model demonstrated superior performance in APT detection compared to existing methods.
• Achieved a precision rate for APT attack prediction between 84% and 91%, an improvement of over 7%.
• The model effectively identified important information and behaviors associated with APT attacks.

Conclusions:

• The BiADG model is a proper and effective method for detecting APTs in network traffic.
• The research validates the superiority and effectiveness of the proposed approach.
• This work opens new avenues for detecting other cyber-attacks like DDoS, botnets, and malware.